Material and methods
Specimens were collected from various parts of Norway in 2019, either from native stands or from cultivated plants of known origin (Figure 1, Table1). Inflorescences were collected and dried for studies of pollen morphology, and herbarium specimens were sampled and prepared for documentation, deposited in herbarium BG.
DNA ploidy levels were measured for the studied specimens except samples of A. rupicola , H. lancifolia , H. neglecta ,H. subpinnata , and S. aucuparia (Table 2). The ploidy levels were measured using flow cytometry analysed by Plant Cytometry Services (Schijndel, The Netherlands) following the procedure given in Andersen et al. (2016).
Anthers were treated with 10% KOH to remove cell content, followed by acetolysis to remove cellulose (Fægri et al., 1989). The pollen was stained using fuchsine and mounted in glycerol. Using light microscope (LM) the polar axis (P) and the equatorial axis (E) of 100 pollen grains of each specimen were measured. The P/E ratio was calculated following Boyd and Dickson (1987). In total 1600 pollen from 16 different specimens were measured. To assess and compare the shape, central tendency, variability of the samples, and to look for outliers in 16 different specimens, the side-by-side boxplots were made of the lengths of the polar axis (P), equatorial axis (E), and the calculated ratio (P/E) of the 100 measured pollen for each specimen (Figure 2).
Scanning Electron Microscope (SEM) was used to investigate the surface structure of the pollen and to aid the description of pollen of the specimens. Pollen was applied with a brush, carefully sprinkled over carbon tape, and glued to an aluminium rivet. Following by coating with a thin layer of gold and palladium to make the sample electrically conductive throughout. Individual pollen grains were visualised with a Zeiss Supra 55VP scanning electron microscope with 5.00 kV, 30 µm aperture with a working distance to the lens ~4 mm. From the resulting pictures, morphological characteristics of the pollen surface were recorded (Table 3).
Principal Component Analysis (PCA) was first run with the three measures pollen characters, the polar axis (P), equatorial (E) axis, and their ratio (P/E). A second analysis with the nine different characters of pollen morphology; including the measured characters and the morphological characters of the pollen surface as given in Table 3. The analyses were performed with centring and standardization of characters and visualized in CANOCO 5 (Ter Braak and Šmilauer, 2012).